topic Re: ?when does marginal performance become unacceptable in Wireless

?when does marginal performance become unacceptable

steve.dutky — Sat, 03 Jul 2021 22:18:26 GMT

Dear all,

I have an issue with a desktop supplier:

The desktop's integral wifi b/g adapter routinely drops about four out a 1000 1000-byte pings from the next-hop gateway through an ap1130 a/g. (for the record, plenty of bandwidth, resources, no problem seen with 3rd party WiFi adapters, usw)

I think this sucks and that he should fix it. He thinks differently.

Can anyone suggest an acceptable threshold for this?

Thanks.

Re: ?when does marginal performance become unacceptable

owillins — Wed, 06 Feb 2008 20:20:22 GMT

Could you please give brief discussion about your topology and features detail.

Re: ?when does marginal performance become unacceptable

steve.dutky — Thu, 07 Feb 2008 22:26:38 GMT

This has a simple sandbox topology:

1130 A/G ap with autonomous IOS trunked at 100/full to a 3550 l3 switchport

I have configured the ap with 3 ssid's, all open authentication, one guest-mode, the other two with 128 bit wep keys.

Each ssid has a corresponding svi/vlan on the switch as a next hop.

The integral adapter drops about 4 in 1000 on the G channel no matter to which ssid it connects.

Please note I don't see any packet drops using other wifi adapters on either A or G frequencies.

Again I'm looking for a rule of thumb. Can you live with losing 4 packets out of a thousand? If yes, how about 1 out of a hundred? If no, how about 4 out of ten thousand?

Thanks,

Re: ?when does marginal performance become unacceptable

mike — Mon, 11 Feb 2008 22:31:19 GMT

4 packets out of a thousdand seems acceptable to me. That's a failure rate of less than half a percent.